@edgescape, sorry to hear of your mother's illness. Prevention is always better than cure, but the science is pretty poor when it comes to prevention when you have risk factors and family history.

The relevant issue for Cogstate is that most trials fail. That's the nature of the work. If Cogstate tools are used, then it doesn't matter if the trial fails because they still get paid.

Multiple failures, however, lead to the issue of the accelerated approvals framework coming under scrutiny, especially in this space.

And that is a material risk for Cogstate, because it's as much a political and philosophical issue (if not more so) as it is about patient safety and science.

@nerdag Thanks for your thoughts. I agree but I think there's always some underlying reason..

The point I made before is not enough is being done to catch the disease early. What is worse is that things such as simply not being able to work a new TV is part of the ageing process when it could be something more serious. Everyone already knows the forgetting names and conversations from yesterday use case is a symptom but by then it is probably too late - they probably already find it difficult learning something new. And probably a prereq for using Cogstate (since it is digital as opposed to paper) as I think Nyck77 stated in an earlier post.

If enough is done detecting the disease early and putting out more awareness, there would possibly be a bigger group which you can pull from to participate in trials which would benefit both Cogstate and the drug companies in this area.

Until that happens through education or awareness, I feel progress will be pretty stagnant. It's frustrating, especially for those who may be invested in this area, not just Cogstate. Thus I think Cogstate or those working on treatments is unfortunately for me in the "too hard" basket for now.

Note that this is my point of view. There are some I know who would see the symptoms and act early but one of the doctors that I talked to says even if found early there's not much you can do.

A larger sample population of earlier stage participants won't materially increase the number of trials if the endpoint being measured with Cogstate's tools (or any other comparable tool) isn't meaningful, or the agents being tested don't really work and pharma loses commercial interest.

Pharma lost interest and pulled pretty much all funding for neuroscience approx 20 years ago before recently barking up the beta amyloid tree. I've made the point before that this could happen again if biologics turn out to be a disappointment as a class.

Cogstate is a trials company first and foremost, and rely on growth in trials for the growth of the company.

If pharma loses interest and pulls the plug, either through failure, or through regulatory changes, that in my view is a material existential risk to Cogstate.

You're right when you say it doesn't change a thing for an affected person unless there is effective treatment. The jury is well and truly still out on whether biologics are going to be that solution.

Cogstate's revenues from treatment or monitoring are negligible, and they will almost certainly not be the only game in town if/when there is an effective treatment (refer to my very old posts on this issue).

I agree that Cogstate is still in the too hard basket for me as a company to invest in.

I agree that large sample population is not going to do much on success of trials because it all depends on the drug.

However, we still need a quality cohort to draw upon when needed and maybe as the disease progresses the cohort member quality probably gets reduced and culled? I think this could be the limiting factor of these trials

Different story though with for example cancer with many participants available and that's why some of the cancer companies such as Telix and Clarity tend to get more traction.

@edgescape, working out what 'n' is required is sufficient for adequate power in a trial is a question of applied statistics.

The question is whether the endpoints picked are valid markers of caseness and improvement in caseness. If the endpoints are weak and meaningless, it doesn't matter how big your 'n' is because the result is, well, meaningless.

Good quality hard endpoints, don't usually need a large 'n' because the study will be sufficiently powered for the study design. Some RCTs will be sufficiently powered with 'n' of 12, others will need hundreds or thousands because of the study design.

Low quality endpoints generally require fancier statistics, or larger numbers, or both. The fancier the applied statistics, the less likely any statistically significant finding will make a meaningful difference.

Hence the controversy about the FDA framework.

@nerdag Actually I'm not really referring to numbers or anything statistical.

When I refer to "cohort quality", I am asking the question of administering drugs and obtain the data to someone that is in cognitive decline.

It's perhaps also a rhetorical question and I don't expect anyone to answer it.

One point we can't argue about is taking care of someone in cognitive decline requires a great deal of empathy where I'm really pushed beyond my "comfort zone"

I think I've discussed enough on this topic of trials and the ability of participants in cognitive decline so will leave it for now.